territory, proximity, and spatiality: the geography of international

Territory, Proximity, and Spatiality: TheGeography of International Conflict1

HARVEY STARR

Department of Political Science, University of South Carolina

This essay presents an overview of the literature on the ways in whichspace, spatiality, and proximity are theoretically important in theexamination of international conflict behavior, for example, to agent-structure models of opportunity, notions concerning diffusion, theloss-of-strength gradient, and the effects of distance and space. Theopportunity and willingness framework is used to organize boththe literature and the discussion, which builds on Diehl’s (1991)seminal overview of geography and conflict.

A recent theme of the annual meeting of the International Studies Association was‘‘The Construction and Cumulation of Knowledge.’’ Both construction and cumu-lation of knowledge are central to the purposes of essays in the ‘‘Reflection, Eval-uation, Integration’’ section of this journal, with cumulation asking scholars to focuson the ways and the extent to which our collective theory and research endeavorhas built upon itself. One approach to cumulation and integration involves bringingtogether the theory and research falling within broad cross-disciplinary categories.That is, scholars will often categorize their work as reflecting a political economy ora political psychology approach or perspective. Another arena that has made asignificant contribution to the study of international relations, but has receivedrelatively less attention, is political geography. This essay is directed toward cumu-lation, reflection, and integration with regard to political geography and the ge-opolitical linkages found in the study of international relations generally andinternational conflict in particular. In doing so, it will explicitly draw on the work ofgeographers and provide linkages between the disciplines of geography and po-litical science.

This essay, then, has several interrelated themes within a broad discussion ofpolitical geography and cumulation. The most prominent is the linkage betweengeography and international conflict. It will be presented within several differentcontexts. First is the more general linkage between geography and internationalpolitics. However, to look at geography and international politics as well as geog-raphy and international conflict, we also need, second, to explicate more clearly aset of concepts that are related to geography and spatiality, including territory,territoriality, proximity and distance, and borders. These concepts help us under-stand the role and importance of spatiality and spatial perspectives and alert us tohow the work of geographers can help us think about international conflict. And,

1This essay is based on a paper presented at the annual meeting of the International Studies Association inPortland, Oregon, February 2003. The author is indebted to the editor and a set of anonymous referees forcomments on an earlier draft of this piece.

r 2005 International Studies Review.PublishedbyBlackwellPublishing,350MainStreet,Malden,MA02148,USA,and9600GarsingtonRoad,OxfordOX42DQ,UK.

International Studies Review (2005) 7, 387–406

third, the linkages between geography and international conflict will be madewithin the context of the opportunity and willingness framework of analysis.

Cumulation and International Conflict

The broad mandate for assessing and possibly improving our knowledge and un-derstanding of the geopolitical aspects of international conflict and internationalrelations touches on a great variety of topics and could lead to an article of immenseproportions. Instead, this essay will comment on several broad approaches to thestudy of geopolitics and international conflict and ways in which to tie studies of thegeography of conflict together as well as present some observations on what wouldbe productive methods for the study of spatial–geographic aspects of internationalconflict. It is hoped that through such an exercise the odds of generating cumu-lationFespecially Dina Zinnes’ (1976) idea of ‘‘integrative cumulation’’ or BruceBueno de Mesquita’s (1989; see also Symposium 1985) focus on ‘‘progress’’ in aLakatosian senseFwill be greatly improved. As noted in Benjamin Most andHarvey Starr (1989:7), for Zinnes, ‘‘additive cumulation occurs when ‘one studyadds some information to the existing literatures on the subject,’ through suchactivities as the citation of previous findings, using previously collected data, sec-ondary or reanalysis of existing data, the incorporation of new cases or new var-iables into the analysis, or expanding the application of models, indices ortechniques to new cases or research questions.’’ Integrative cumulation goes fur-ther, reflecting instances in which earlier studies are ‘‘ ‘crucial’ to the conceptual andtheoretical components of the subsequent study’s research design.’’ Additive cu-mulation should not be slighted, however, in that it is an indispensable part of theprocess that leads to integrative cumulation.

Territoriality, proximity, and spatiality have all played central roles in the study ofinternational conflict. The place of territory in the analysis of conflict and the studyof the diffusion of conflict are two areas in which there has been considerable‘‘additive cumulation.’’ This is especially true regarding the use of border data setssuch as the extensive and detailed data set compiled by the Correlates of Warproject (whose homepage is now located at Pennsylvania State University at http://cow2.la.psu.edu/). Additive cumulation is also found in the widespread use of con-tiguity as an independent or control variable in the study of conflict across differentunits of analysis (monadic, dyadic, and regional analyses), different time periods,and for different types of conflict (for example, war only or events lower on theescalatory ladder as tapped by the Militarized Interstate Dispute or MID data). Theconcept of ‘‘nice laws’’ (see Most and Starr 1989)For domain specific lawsFsug-gests that we should be concerned with under what conditions certain theorieshold. In doing so, activities like those noted above, which indicate additive cumu-lation, become important and useful in specifying theory.

Given the extensive set of factors in the area of political geography that have ledto additive cumulation, it is important to note that examination of the place ofterritory in the analysis of conflict and the study of the diffusion of conflict have alsodemonstrated a growing ‘‘integrative cumulation.’’2 The argument for cumulationin these areas (contra Simowitz 1998) has been presented in a set of overview workson the role of territory (which will be discussed below) and of diffusion studies (forexample, see Most, Starr, and Siverson 1989). Indeed, except for the researchprogram on the democratic peace (see Russett and Starr 2000), few other areas

2In lieu of an extensive discussion of cumulation and its role in the research programs of international relationsscholars, readers are directed to Zinnes’ (1976) seminal chapter. See also Most and Starr (1989: chapter 1) as well asStarr (2002a, c) for further discussions of cumulation as it relates to the research triad of theory, logic, and research

design in general. See Most, Starr, and Siverson (1989), Starr and Siverson (1998), as well as Kristian Gleditsch(2002) for discussions of cumulation as it applies specifically to the diffusion of conflict.

Territory, Proximity, and Spatiality388

in the study of conflict have developed such consistent, complementary, andreinforcing sets of empirical findings (see, for example, Hammarstrom and Heldt2002). As noted, variables representing proximity or the issue of territory are rou-tinely included in analyses as standard components, either as controls or as specificindependent or intervening variables. In effect, contiguity has become a standardcomponent of models since its inclusion as one of the handful of factors in StuartBremer’s (1992) famous identification of ‘‘dangerous dyads.’’ How territory, prox-imity, and spatiality have promoted integrative cumulation in the study of inter-national conflict will be addressed in the remainder of this essay.

Spatiality and International Conflict

But first, key concepts need to be explicated. Students of international relationshave developed unusually coherent and useful overviews of theoretical and em-pirical work that relate the study of international politics and conflict to geography,territory and territoriality, distance, space, and spatiality. Even though the over-arching idea that holds all of these works together is that of the ‘‘spatiality’’ ofphenomena, overviews of spatiality by scholars who are not geographers have ap-peared only relatively recently. The author (Starr 2001a, 2003) has begun such aprocess by reviewing the thinking about space and spatiality by geographers to helppolitical scientists in understanding why the spatial or locational contexts of be-haviors need to be included in our studies along with the temporal contexts.3 Indoing so, it became clear that we need to understand more fully the differencesbetween such concepts as ‘‘space,’’ ‘‘location,’’ and ‘‘distance.’’

Let us begin by drawing upon approaches to the idea of ‘‘place’’ by geographers(for example, Agnew and Duncan 1989):

The first approachFspaceFis the one with which . . . students of conflict are mostfamiliar, emphasizing the location of things in relationship to other things, and howthings are distributed. This idea of spatial contingency is picked up in Kirby andWard’s (1987:3) definition of ‘‘spatiality’’ as ‘‘a contingent factor within the op-eration of any social formation,’’ in which society’s ‘‘components are themselvesdependent upon their spatial setting’’ (see also O’Loughlin and Anselin 1992:12).. . . This view of spatiality or location matches the two basic ways to think aboutlocation, as presented by Abler, Adams, and Gould in their classic text, SpatialOrganization (1971:59)F‘‘absolute location’’ and ‘‘relative location.’’ According toAbler et al., ‘‘absolute location is position in relation to a conventional grid systemdesigned solely for locative purposes.’’ In this view, location is provided by suchthings as latitude and longitude or a street address. The concept becomes muchmore rich and trickier in the second way to think about location: ‘‘Relative lo-cation is position with respect to other locations.’’ This can be expressed in termsof distance or travel times from other locations, the cost of such travel, etc. Thus. . . technology changes ‘‘relative location’’Fplaces that were once weeks apart intime are now only hours apartFand other mechanisms, such as alliances, can dolikewise (Starr 2003:5, emphasis added).

As can be seen, with spatiality we are confronted with the question of ‘‘dis-tance’’Fhow close or far units are within some concept of space. Abler, Adams, andGould (1971:72) have noted that prior to 1950 geographers most often dealt withspace (and distance) in the ‘‘absolute’’ mode. Generally since then, in most research‘‘relative location and relative distance [have] been used to define new kinds ofstretchable, shrinkable spaces.’’ They (Abler, Adams, and Gould 1971:82) go on to

3For other approaches to spatial issues, see the work of geographers such as John O’Loughlin and Luc Anselin(1991, 1992), the 2002 special issue of Political Geography edited by Michael Ward as well as work on related concepts

such as ‘‘landscapes’’ (see Axelrod and Bennett 1993). An especially useful piece on geospatial ontology is that byAntony Galton (2001).

HARVEY STARR 389

observe that ‘‘human decisions constantly alter and restructure relative spaces. Ithas taken geographers a long time to challenge the pervasive tyranny of absolutespace.’’ Not only do human decisions ‘‘alter’’ relative spaces, various types of rel-ative space explicitly take ‘‘time’’ into account, so that relative space and relativedistance (and the meaning of relative space or distance) are heavily dependent onperceptions. As Abler, Adams, and Gould (1971:75) argue, ‘‘the spaces in whichpeople live are much more psychological than absolute.’’

The concept of relative distance thus complicates the question of how close or farthings are from one another. This is a classic question, raised by many students ofconflict, including in Kenneth Boulding’s (1962) seminal notion of the loss-of-strength gradient (LSG), which has been used in many subsequent studies (forexample, Bueno de Mesquita 1981; Lemke 1995). In setting out the necessaryelements of a conflict, Boulding discussed the need for at least two ‘‘parties’’ andimportant characteristics of these parties. One was the general principle ‘‘that eachparty can be supposed to be at his maximum power at home . . . but that hiscompetitive power, in the sense of his ability to dominate another, declines thefarther from home he operates. This is the great principle of the further the weaker’’(Boulding 1962:78–79, emphasis in original). The loss-of-strength gradient is theamount of competitive power that is lost per some unit of distance from home.Here, relative distance is measured, not through some absolute distance, but ametric in the decline of power.

The question of ‘‘distance’’ was also raised by Quincy Wright (1942). Inhis massive work searching for the causes of war, Wright (1942:1240) hypothesizedthat the greater the ‘‘distance’’ between states, the greater the probability ofwarF‘‘when powers are so isolated from one another that there is no basis formutual understanding.’’ He attempted to measure distance through the manyforms of physical and psychological distances that exist between social unitsFtech-nological and strategic, intellectual and legal, social and political, psychic and ex-pectancy, and policy distances. A similar notion of distance was also the basis ofBruce Russett’s (1967) study of international ‘‘regions,’’ in which he clustered statesin terms of their distances on variables such as trade, UN voting, religion, andculture. Two recent reviews of ways to think about and measure distance are thoseby Gleditsch and Ward (2001) and Alan Henrikson (2002). Whereas Henriksonreviews gravity models of distance along with topological and attributional modelsof distance (such as those used by Wright and Russett), Gleditsch and Ward(2001:744) present a way to go beyond the either/or measurement of contiguity andreconceptualize distance by ‘‘measuring the shortest distance between the twoclosest physical locations’’ for any pair of states.

Returning to the issue of integrative cumulation and how theory develops, it canbe argued that students of international relations have been concerned with dis-tance for two broad theoretical and conceptual reasons. Conveniently, these tworeasons can be summarized as ‘‘opportunity’’ and ‘‘willingness.’’ First, distance isimportant because states (or any other social units) that are close to each other, thatis, are in proximity to one another, are better able to interact. Simply, they have thepossibility or opportunity of interacting with one another. This is the ‘‘interactionopportunity’’ argument or approach to be discussed below. It derives directly fromthe work of Harold and Margaret Sprout on ‘‘environmental possibilism’’ (see Starr1978; Most and Starr 1989: chap. 2). The Sprouts were concerned with what theycalled the ‘‘ecological triad,’’ which is composed of three elements: (1) an actor, orentity, of some sort, (2) an environment that surrounds the entity, and (3) the entity-environment relationship. In searching for alternatives to older geopolitical modelsof determinismFin which, by definition, decision makers are incapable of choicegiven the characteristics of the geographical environment or ‘‘milieu’’ (Sprout andSprout 1969:44)Fthe Sprouts identified ‘‘environmental possibilism,’’ referring tothe set of factors that limit human opportunities and constrain the type of action


that can be taken as well as the consequences of such action (Most and Starr1989:27). The environment presents policymakers with a range of possibilities,which, according to the Sprouts, had to be discoverable by the decision-makingentity.

Willingness refers to ‘‘the choice (and process of choice) that is related to theselection of some behavioral option from a range of alternatives. Willingness there-fore refers to the willingness to choose (even if the choice is no action), and toemploy available capabilities to further some policy option over others’’ (Most andStarr 1989:23). This concept is closely related to the Sprouts’ notion of ‘‘cognitivebehaviorism,’’ another alternative to determinism in the study of the ecologicaltriad: ‘‘the simple and familiar principle that a person reacts to his milieu as heapperceives itFthat is, as he perceives and interprets it in light of past experience’’(Sprout and Sprout 1969:45). Thus, the second reason why we should be con-cerned with distance is because states (or any other social units) that are close toeach other are also perceived as important or salient to each other. This is so for acombination of reasons. Greater perceptions of threat or gain, or of interdepend-ence, are ways in which proximity can generate salience. States (or whatever unit isunder consideration) that are close are seen as more important. Such views affectwillingness through the expected utility calculations of policymakers. Willingness tointeract and to manage subsequent conflicts in different ways, for example, willdepend on the importance or salience of an issue or an opponent.

Thus, proximity makes states (or other social units) that are close to one another‘‘relevant’’ to one another through some combination of both opportunity andwillingness. High levels of opportunity and willingnessFgenerated, for instance, bylong contiguous borders that go through areas with valuable resources, importantstrategic features, and on both sides of which live members of the same ethnicgroupFmean that two states are both easily able to interact with each other andboth perceive the other as important and relevant (whether as a possible opponentor cooperator through shared interests).

This presentation becomes more important if we look at the way many researchdesigns to study international conflict are constructed. Students of internationalconflict have structured research designs to include only ‘‘relevant’’ dyadsFpairs ofstates that are able to interact with one another, highly likely to interact with oneanother, or perceive important stakes involved in that interaction (for example,Lemke 1995; Leeds and Davis 1999; Lemke and Reed 2001). They have developedstudies based on states within politically relevant areas or neighborhoods (see, forexample, Maoz’s ‘‘politically relevant international environment’’ or PRIE, 1996;Enterline 1998; Murdoch and Sandler 2002). New work on ‘‘network’’ analyses ofvarious kinds (for instance, Maoz 2001, 2002; Gleditsch’s 2002 ‘‘connectivity ma-trix’’ analysis; Hammarstrom and Heldt 2002) extends the concept and utility ofactors who are ‘‘relevant’’ to each other through spatial or behavioral proximity. AsMats Hammarstrom and Birger Heldt (2002:358) explain, ‘‘the term ‘network’refers to a set of units of some kind and the ‘ties’ (relations) of specific types thatoccur among them.’’ Not only can such ties or relations be represented spatially, butthey are often dependent on the actual spatial arrangement of the social units. Forexample, using ‘‘network position,’’ Hammarstrom and Heldt (2002) support pre-vious findings on the diffusion of conflict among contiguous states, but they alsoargue that the network position approach allows them to identify which contiguousstates are most likely to be the sites of diffusion.

Now Add Territory

There is an additional factor in the relationship between proximity and the stakes ofinteraction. If we are talking about states, we are talking about ‘‘territorial units.’’States are proximate to one another in a spatial or geographic manner if their

HARVEY STARR 391

territorial areas are near each other. How close or far are these territorial areasfrom each other in terms of absolute distance? Are they contiguous? That is, do theterritories of two states ‘‘touch’’ each other? Do they ‘‘border’’ each other? If theydo not actually touch each other, are they separated by rivers? If they do notactually touch each other, how far apart are they across some other body of water?Thus, borders represent the highest level of proximityFthe touching of territoryor, by dictionary definition, the condition of contiguity (see Starr and Most 1976;Gochman 1992 for conceptualizations of borders).

It should be clear that territory serves at least two distinct purposes in the studyof international relations. First, by defining the territorial extent of political units,territory creates spatial arrangements among the units indicating the physical–geographic distance between those units. This ‘‘distance’’ is dynamic, in that the‘‘time–distance’’ between the units changes with changing technologies of trans-portation and communication, with changes in the arrangements of the unitsthrough alliances, or with the merging of units through conquest or voluntaryintegration. As noted in Starr and Siverson (1990), alliances can act as ‘‘politicaltechnology’’ by changing not only the absolute distance between units, but theirtime–distance as well. State A, which does not have a contiguous border with StateB, is now able to border State B through an alliance with State C, which is con-tiguous to B. This exact issue arose among NATO, Turkey, and Iraq prior to andimmediately after the onset of the Second Gulf War, given the desire of the UnitedStates to reduce the loss-of-strength gradient of bringing military force to bear onIraq. The conquest of territory can create new borders (as in the expansion ofRussia or the United States across their respective continents or the Napoleonicexpansion across Europe). The breakup of states or empires can create new bor-dersFas in the post-Soviet Union or post-Yugoslavia situations or the redrawing ofthe world map after World War I with the dissolution of the Ottoman, Russian, andAustro-Hungarian empires. But borders were also altered with the reunification ofGermany after the fall of the Berlin Wall and the dissolution of the Warsaw Pact.Voluntary integration, as seen in the progressive growth of the European Union(both geographically and functionally), can also change the legal, economic, andpolitical nature and meaning of borders.

Second, as the place where people live, territory provides an important componentof ‘‘group identity’’ and becomes endowed with extraordinary symbolic importanceto people. There is a large and growing literature on identity and geography, par-ticularly as generated by geographers concerned with the symbolic importance in-herent in territory and the role such symbolism and identity plays in the daily lives ofpeople as well as international politics (see, for example, Agnew and Duncan 1989;Agnew and Corbridge 1995; Newman 1999; Kliot and Newman 2000; Dodds 2001).In addition to value based on symbol and identity, territory may also provide realresource value to peoples (arable land; potable water; minerals of value such as gold,uranium, or oil; access to seas or rivers; and other features of militarily strategicvalue). So, as will be noted below, territory takes on value across many dimensions:territory is important to humans across all levels of social aggregation. It both be-comes a source of conflict and raises the stakes of any conflict.

These broad ways in which territory becomes related to conflictFagain throughopportunity (ease of interaction) and willingness (importance or salience)Fhavebeen addressed by a number of scholars in literature reviews or as part of specificresearch projects. Indeed, a special issue of International Interactions and the follow-on edited volume (Ward 1992) contain a number of important articles linkingdistance-proximity and territory-territoriality to international conflict, includingthat by O’Loughlin and Anselin (1991), which argues how geographers need tohelp bring geographic elements back into the study of international relations. Per-haps the most influential of these pieces was by Paul Diehl (1991); it has served asthe basis for subsequent reviews and empirical research with scholars building on


his broad categorization of the two basic ways territory becomes part of the inter-national conflict process (to be discussed below).

This collection was soon followed by a book by Gary Goertz and Diehl (1992) onterritorial change and conflict, significant sections of John Vasquez’s (1993) WarPuzzle, and Paul Huth’s (1996) work on territorial disputes. A more recent collec-tion on geopolitical and territorial studies was also edited by Diehl (1999; seeespecially Diehl’s introductory chapter). An important review article on the rela-tionship between territory and conflict, drawing on his own recent work and fol-lowing on themes developed by Vasquez and Diehl, is found in Paul Hensel (2000).

The Geographic Context

Territory is the most obvious way to highlight the importance of geography. Butwhen we take space and spatiality, proximity and distance, and territory together,we are identifying the need that social relations be studied within a geographic orspatial context as well as a temporal context. Many of the literature reviews andedited volumes noted above deal with, and stress, the importance of the spatialcontext (see also Goertz 1992, 1994; Starr 2001a, 2003). The relevance of space,proximity-distance, and territory is evident in Goertz’s (1992:301) definition ofcontext: ‘‘factors that influence fundamental relationships and meanings.’’ Goertzhas also stressed the crucial ways in which context affects research design. Considerthat the use of time series data and designs is becoming more standard and the useof time to delineate units of analysis (for example, ‘‘nation-year’’) is currently thenorm. In these ways, despite some exceptions (the study of contiguity, for example),much research in international relations involves the temporal dimension or tem-poral context.

Even though there is no denying the significance of time and the temporalcontext in social science, analysts need to pay more explicit and extensive attentionto the spatial elements, or the spatial contexts, of social phenomena. Why? It isbecoming clearer that time and space are two of the primary ways in which wecontextualize social behavior and interactions. Abler, Adams, and Gould (1971:10)succinctly indicate the importance of these two dimensions: ‘‘Time and space areobvious and immediate aspects of human existence. . . . Time and space are thefundamental contexts of all experience. . . . Experience must be located in time andspace before we can begin to process it further. . . . Locating an event in the spatio-temporal continuum is our first step in ordering our experience of it.’’ Althoughthis argument is rather obvious, it is crucial to understanding why any form ofcumulation needs to be built around space as well as time. And it is key to un-derstanding that analyses structured solely (or almost entirely) around time areonly telling us half of the story. The welcome appearance of studies linking thespatial context to international conflict (as well as the synthesizing literature reviewsalready noted) have gone a long way to helping us understand the ‘‘other half ’’ ofthe story. Studies of the importance of territory and the effects of changes in ter-ritory on international conflict demonstrate how information about the territorial-spatial context is clearly needed to complement and fill out the temporal contextfound in studies, for example, of enduring rivalries (see the works of Diehl andGoertzFboth together and separatelyFmentioned above; also see Vasquez 1996;Rasler and Thompson 2000).

Opportunity, Willingness, and Ordering Spatial-Geographic Effects

We have already shown above how the opportunity and willingness framework canhelp represent the importance of distance as well as the effects of territory ininternational politics. We will now use opportunity and willingness to furtherexplore how territory-geography is related to conflict. But in so doing, we will also

HARVEY STARR 393

be returning to the initial development of opportunity and willingness as a way tohelp synthesize and bring order to large (and sometimes unwieldy) literatures (seeStarr 1978).

Looking across the literature on territoriality, geography, and international con-flict, several broad themes emerge. As with the general study of conflict, the con-cepts of opportunity and willingness can be useful in organizing literatures andmaking sense of disparate studies and approaches as well as can serve minimally asa pre-theoretic device for generating hypotheses, conceptualizing components ofour theories or models, and searching for nonintuitive relationships (see Starr1978; Most and Starr 1989; Cioffi-Revilla and Starr 1995). Although these ideas willbe familiar to many readers (now used so frequently that many authors do not feelit necessary to supply citation or attribution), a further (although still brief ) over-view is in order.

Opportunity and willingness is a form of agent-structure model, initially createdto deal with the ways in which ‘‘entities’’ are related to their environments. To goback to basics, interdependence is a quality of systems. Systems are composed ofunits of some kind and the interaction among them. In the simplest of terms, wemust be concerned with each unit and how each unit adapts to its environment.This individual adaptation produces the patterns of interaction that characterizethe system. The ecological triad of the Sprouts described above informs any agent-structure approach and helps us think about units and their environments.

The seemingly elemental construct of the ecological triad has served as the basisfor the development of the opportunity and willingness framework. The ecologicaltriad provides great utility in its ability to link the entity and the environment byhelping us see how and why different environments constrain, limit, or enable whatentities are able to do and what they are likely to do. As critics of deterministicgeopolitical or environmental models, the Sprouts presented alternative forms ofthe entity–environment relationship. Three of these are particularly useful to ourthinking; two of which were noted above. First is ‘‘environmental possibilism’’ inwhich the environment is seen as a set of constraints on what is actually possible forthe entity to do in that particular environment. Second is ‘‘environmental pro-babilism’’ in which the environmental constraints and possibilities make certainbehaviors more or less likely. Third is ‘‘cognitive behaviorism’’ in which the en-tityFas ultimately embodied by individual decision makersFis linked to the en-vironment through the images of the environment that people hold. Ultimately,then, we are concerned with the possibilities and constraints that face decisionmakers (opportunity) and with the choices that they make in light of these pos-sibilities and constraints (willingness). The various levels of analysis involved in thestudy of international relations are thus linked by thinking of a decision maker as anentity who must behave within the very complex environment that surrounds himor her. Each level of analysis used in the exploration of international politics andforeign policy (for example, idiosyncratic-psychological, role-organizational, gov-ernmental, societal, dyadic or relational, regional, systemic) describes one of theenvironments within which the decision maker must operate (see Gleditsch 2002for an application of this framework to zones of regional clustering).

The environments in which decision-making entities operate provide opportu-nities, risks, and potential costs and benefits that constrain policymakers. How areall these elements captured by the concept of opportunity? The environment makescertain opportunities, and not others, possible. Here the environment is seen as aset of constraints on what is actually possible for the entity to do in that particularcontext. Goertz (1994) has elaborated on this notion in his discussion of context as‘‘barrier.’’ Possibility includes two dimensions. First, the phenomenon must alreadyexist somewhere in the world system. The phenomenonFbe it nuclear weapons,telecommunications satellites, Protestantism, Marxism, railroads, or financialmarketsFmust have been ‘‘invented’’ so that it is available as a possibility to at


least some actors in the system. The second dimension centers around the distri-bution of this possibility in the system. For example, nuclear weapons do exist;however, most states cannot ‘‘take advantage’’ of them because they have neitherthe wealth nor the expertise to produce their own. Though a possibility may exist,limits on resources affect the ability to make use of it.

To summarize, opportunity requires three related conditions: (1) an environ-ment that permits interaction between states, (2) states that possess adequate re-sources to be capable of certain kinds of actions, and (3) decision makers, or humanagents of some kind, who are aware of both the range of interactions and the extentof capabilities available to them. Opportunity is the possibility of interaction becauseof objective conditions that may be perceived in varying (more and less accurate)ways by decision makers.

Willingness is concerned with the motivations that lead people to avail themselvesof opportunities. Willingness deals with the goals and motivations of decision mak-ers and focuses on why decision makers choose one course over another. Willing-ness is therefore based on perceptions of the global scene and of domestic politicalconditions. It derives from calculations of the costs and benefits of alternativecourses of action, based not only on objective factors but also on perceptions (forinstance, of threat) and emotions (for instance, fear, insecurity, or desire for re-venge). Willingness thus depends on choice and perception. A person reacts ac-cording to what she thinks she can do and what others expect her to do. Willingnessinvolves all those factors that affect how decision makers see the world, processinformation about the world, and make choices.

Finally, it is important to understand that both opportunity and willingnessare required for a given behavior to occur; they are jointly necessary conditions.Wishing for something to happen is not enoughFthe capabilities to act for itsfulfillment must be available. Simply being able to do something does not mean itwill happen unless you have the will to take action (see especially Cioffi-Revilla andStarr 1995).

Geography as a Facilitating Condition

Let us now return to the relationships between geography and conflict as well asterritory and conflict. Recall that territory reflects the spatial location of states,including their proximity or distance from one anotherFwith their contiguousborders representing the highest level of proximity. We have also indicated some ofthe reasons why proximity or distance and territory are important. We can returnto these basic issues using Diehl’s (1991) seminal article assessing the work ongeography and war (see also Goertz and Diehl 1992:chap. 1). This work is impor-tant because of the manner in which Diehl categorizes the literature. He breaks theempirical studies of geography and war into two groups: (1) ‘‘geography as a fa-cilitating condition’’ for conflict, and (2) ‘‘geography as a source of conflict’’FinGoertz and Diehl (1992), the word ‘‘geography’’ is replaced by ‘‘territory’’ but thearguments remain substantially the same. As noted above, here we are placingDiehl’s analysis into a broader framework by using opportunity and willingness tohelp indicate how the analyses of geography-territory and war reflect the same sortsof phenomena highlighted when applying opportunity and willingness to distanceand territory.

The distinction that Diehl presented between ‘‘geography as a facilitating con-dition’’ for conflict and ‘‘geography as a source of conflict’’ has been influential instudies relating territory and geopolitical factors to conflict. For instance, Hensel(2000) speaks of territory as ‘‘context’’ (that is, as a facilitating condition) and as a‘‘source’’ of conflict. Although Vasquez (1996) has provided three theoretical per-spectives linking geography and conflict, they also reduce to Diehl’s two views.Vasquez’s ‘‘territoriality perspective’’ discusses geography as a source of conflict,

HARVEY STARR 395

something that states fight over. His two other perspectivesF‘‘proximity perspec-tive’’ and ‘‘interaction perspective’’Ffall under Diehl’s rubric of geography as a‘‘facilitating condition.’’ Vasquez’s proximity perspective is about the ease withwhich states can reach each other militarily and the interaction perspective linkscloseness or proximity to the frequency of interaction.

The work of Starr and Most (1976) specifically looked at the importance ofproximity, and how it was measured and represented by borders and territory.Their diffusion research project (Most and Starr 1980) moved to examine bordersafter concluding that the diffusion of certain phenomena was best studied by look-ing only at units that were ‘‘relevant’’ to one another. Such relevance could beindicated by geographical proximity. Proximity, in turn, could be operationalizedthrough ‘‘borders,’’ which were viewed as important indicators of proximity be-cause they exhibit important relationships with both the opportunity and willing-ness of state actors as conceptualized by Starr and Most (1976).

Indeed, one set of studies Diehl used to illustrate the category of ‘‘geography as afacilitating condition’’ for conflict is the combined work of Starr, Most, and Siverson(for example, Most and Starr 1980, 1989; Siverson and Starr 1991). In this re-search, Most and Starr develop the idea of ‘‘interaction opportunity’’ based on thework of Boulding as well as a number of geographers. Boulding’s (1962) LSG isbased on the greater costs and the greater difficulties of operating militarily, thegreater the distance the target is from a state’s home territory. Earlier, the geog-rapher G. K. Zipf (1949) presented his ‘‘law of least effort,’’ which simply arguedthat units interact more with those that are closer to them and less with thosefarther away. In Most and Starr’s terms, the closer units such as states are to oneanother, the greater their possibility for interaction.

The opportunity for interaction concept was elaborated by Siverson and Starr(1991). They conceptualized proximity, as measured by borders and contiguity, as afactor of ‘‘loose necessity.’’ That is, proximity creates the possibility for conflictthrough increased possibilities for interaction (both positive and negative); thus, itraises the probability of interactions, both positive and negative. They stress thebasic idea that the interaction opportunity model only holds that closer units willinteract more.

However, many (if not most) scholars have assumed that greater interaction leadsto more conflict and, in turn, their studies have been designed to apply the in-teraction opportunity idea to conflict rather than cooperative behavior. Such anapproach is clearly a misunderstanding of possibilism as well as opportunity. Thegeographer Zipf ’s (1949) ‘‘law of least effort’’ applies to interactions in general,positive as well as negative. Diehl (1991) understood this distinction, noting thatconflict was not the ‘‘necessary’’ effect of having an interaction opportunity throughcontiguityFmeaning that proximity of territory only increases the probability ofconflictual interaction but does not ensure it will happen.

Likewise, Siverson and Starr (1991) found that borders (as well as alliances)only increase the probability that ongoing wars will diffuse to ‘‘warring bordernations,’’ not that they necessarily will do so. Similar results abound. For example,Maoz and Russett (1992:260) have observed that even though contiguity is astrong factor in predicting dyadic conflict, ‘‘it does not account for the relativelack of conflict between democratic states.’’ And, of course, none of the argumentsfor the interaction opportunity apply only to territorially proximate homelands.Starr and Most (1976) raise the issue that territorially proximate possessionsof states will have the same effects. In addition, the interaction opportunity argu-ment (based on the Loss of Strength Gradient) also recognizes that ‘‘great’’ or‘‘major’’ powers are so named because they possess a greater ability to interactwith states far from their homelands. These states can project military powerglobally so that their interaction opportunities have transcended first-order terri-torial contiguity (see, for example, Morton and Starr 2001 for a discussion of ‘‘elite


powers’’ that combine large land armies with large navies or strategic bombers andmissiles).

Indeed, recent studies have shown that high interaction opportunities may leadto more cooperative behavior as well as conflictual behavior. Starr and Thomas(2002a, 2002b, 2005) have found that high levels of ease of interaction acrossbordersFgreater interaction opportunitiesFare also related to positive De-utschian interdependence-integration effects. They argue (Starr and Thomas2005) that simply categorizing two states as being contiguous may not adequatelyreflect expected underlying behavior. And they present two different views on therelationship between contiguity and conflict: (1) interaction opportunity, which ishypothesized to make conflict more probable, and (2) Deutsch’s social communi-cation model of integration, which contends that increased interactions, transac-tions, and interdependence make conflict less probable. Each view, however,represents a linear (positive or negative) relationship between ease of interactionand conflict. To deal with this problem, Starr and Thomas (2002b, 2005) propose acurvilinear relationship with the low occurrence of conflict at both the lowest andhighest levels of ease of interaction (opportunity) and salience (willingness). Conflictis most likely when the expected utility of conflict is greatestFthat is, in the middle,where states have both the opportunity and willingness to engage in conflict.

These observations stress the importance that those studying territory and prox-imity understand the arguments behind interaction opportunity and ease of inter-action as a facilitating condition correctly. As a form of opportunity, the facilitatingcondition argument starts with the idea that it must be possible to interact, to haveconflict, and to have militarized conflict. Proximity both creates such possibilitiesand raises their probabilities (and also raises the probability of cooperative inter-actions under the right circumstances). But, proximity is only one of a number ofother potentially substitutable ways by which these possibilities occur! Most andStarr (1989) have observed that opportunity or willingness can operationally occuror be made available in a number of alternative, nonunique, and substitutable ways.Substitutability, then, refers to the existence of a set of alternative modes of re-sponse or ‘‘alternative modes of redundancy’’ (Cioffi-Revilla and Starr 1995:456–457) by which decision makers can deal with some situation. Again, proximity is justone of the substitutable modes to increase interaction opportunities. Any factor thataffects the meaning of distance, especially time–distance, can become such a sub-stitutable mechanism. Technology provides many such mechanisms. For example,possibilities for interaction with a power projection that is low in cost exist for stateswith long-range nuclear-armed missiles or aircraft carrier-based bombers.

As noted above, behavior cannot occur without both opportunity and willingness.Opportunity can be created by a number of ‘‘second-order substitutable mecha-nisms’’ (the terminology used by Cioffi-Revilla and Starr 1995). Spatial proximity(for example, through contiguity) is one such mechanism. Thus, geography orterritoriality as a facilitating mechanism is not a contending model with geography–territory as a source of conflict. As the two models represent opportunity andwillingness, both must be present. The facilitating condition makes conflict possible,increases its probability, but does not guarantee that it will occur.

Geography as a Source of Conflict

The work on geography, territory, or the ‘‘territorial perspective’’Fin which ter-ritory acts as the cause, the source, or the stakes involved in conflictFis quiteextensive. Diehl (1999: x) reviews the value or importance of territory or what hecalls the tangible ‘‘intrinsic importance of territory’’ including such items as naturalresources, control over populations, access to trade, and strategic value. More in-tangible or symbolic aspects of territory are also included such as its historic valueand its relationship to the group identity–ethnicity of the people living on it. David

HARVEY STARR 397

Newman (1999:14) sees territory as a ‘‘demographic container’’ that holds people,providing territorial symbolism to their identity, such that territory becomes an‘‘exclusive entity’’ for a people. This view creates a powerful we–they or us–themdivide regarding territory (see, for example, Henderson 1997) and is vital to mod-els based on the cohesion of social groups (for an overview of social-psychologicaltheories see Pruitt and Kim 2004). For instance, Cameron Thies (2001) has linkedterritory as an issue to the factor of national identity, discussing how this connectionhelps generate and maintain an enduring rivalry.

Given the variety of factors that give territory value, it is not surprising thatterritory serves as the origin of conflict. It does so through territorial claims (Hensel2001), territorial changes (Kacowicz 1994), territorial disputes and their settlement(Huth 1996; Gibler and Vasquez 1998; Hassner and Hironaka 2002; Huth andAllee 2002; Sample 2002), and strategic concerns, among others. Huth (1996) hasused territorial disputes to demonstrate how domestic factors interact with suchdisputes to modify realist geopolitical theories. An interesting twist on the strategicimportance of territory is found in John Vanzo (1999), who is concerned with theconfiguration of borders, especially the notion of the ‘‘compactness’’ of states.

Arguments for, and reviews of, empirical findings regarding territory-geographyas a source of conflict can be found in Kal Holsti (1991), Goertz and Diehl (1992),Vasquez (1993), Stephen Kocs (1995), Douglas Gibler (1996), Hensel (1996, 2000),and Huth (1996). One illustrative empirical study is that by Vasquez and PaulSenese (2003), which indicates that territorial claims increase the probability that apair of states will engage in a militarized dispute (MID) and that such territorialMIDs, in turn, increase the probability of war. A review of the continuously updatedset of contemporary armed conflicts by Peter Wallensteen and his colleagues in theJournal of Peace Research also reveals the extent to which territorial issues continue tounderlie international conflict in the post-Cold war era: 134 of the 225 ‘‘armedconflicts’’ identified between 1946 and 2001 (or 60 percent) involved territory asthe ‘‘incompatibility’’ in the conflict (see the data sets listed at www.prio.no/page/preview/preview/9429/40484.html). Similarly, sixteen of the twenty-nine conflictsfound in 2003, or 55 percent, were based around territorial incompatibilities(Eriksson and Wallensteen 2004).

Territory (or geography) as a source of conflict falls under both opportunity andwillingness. Territory that connects, sits between, or is disputed by two states pro-vides something to fight over. This is not simply a facilitating condition. Territoryexists as a possible issue for conflict; it is available as a source of conflict or con-tention. Because of territory’s tangible and intangible value, it is something thatpeople care about and are willing to fight over. Thus, territory is directly connectedto willingness. People, groups, and states come into conflict every day duringthe course of normal social transactions and interactions: incompatibilities occur,representing incompatible claims of interests and preferences. Most such incom-patibilities are managed simply through routine mechanisms or are ignored be-cause they do not make claims to things that are highly valued. The researchon territory and conflict indicates that territory is literally always of high value,salience, or importance to people and groups. Territory raises the stakes or valueof conflict, thus raising the probability of escalation and lowering the probabilityof easy management. In other words, because territory is of such value, it increasesthe expected utility of fighting for it, even if the probability of success appears tobe low.

Perhaps the study of protracted social conflicts best represents all of these points(see Azar 1984, 1985; Friedman 1999, 2002). In Gil Friedman’s (2002:11) terms,the typical protracted conflict situation finds the ‘‘geodemographic integration ofrival nations,’’ that is, the intermingling of peoples from different ethnic-nationalgroups on the same territory. This constant opportunity for conflict is also em-bedded within a context of constant willingness. The conflicting claims over the


ownership of the territory are claims representing the highest values for eachgroup because the ownership of the territory is passionately attached to or em-bedded in group identity. The constant, and high, level of both opportunity andwillingness is a significant factor in the ‘‘intractability’’ (Azar 1985) of protractedsocial conflicts.

It follows that the decision regarding conflict or escalationFwillingnessFismade more probable for issues involving territory because the value of territorycontributes to the positive expected utility of choosing conflict or conflict escalation.Hensel (2000), for example, explicitly links territory to such expected utility con-siderations. This is a key reason why territory and contiguity are regularly usedas independent, intervening, or control variables in models of international conflict.It is important to note that, just as with territory as a facilitating condition, theargument here is not that territory is the only, or even the most important, sourceof conflict across all situations or under all conditions (see, for example, Mitchelland Prins 1999 who demonstrate this observation while investigating the issuesat stake in the study of militarized disputes). What the work of Diehl, Vasquez,Hensel, Huth, and others mentioned earlier, however, has shown is that territory isoften the central issue at stake and should be included in the group of ‘‘usualsuspects.’’

The relationship between the value of territory and international conflict can alsobe explained by prospect theory (or, more specifically, the prospect theory variationof expected utility models; see, for example, Levy 2000). Prospect theory intro-duces the idea of ‘‘endowment effects,’’ in which, ‘‘because of loss aversion, peopletend to value what they have more than comparable things that they do not have’’(Levy 2000:195). The new acquisition of territory (because it is so highly valued)produces almost immediate endowment effects. Because of these endowment ef-fects when a country takes territory, we find that both sides now frame the situationas one of losses. That is, the state that loses territory frames the situation as one ofloss, becoming risk acceptant in terms of the escalation or pursuit of the conflict.And, in turn, the state that has newly acquired territory now claims the territoryand also quickly frames any return of the territory in the realm of losses. Thus, bothsides frame the situation as involving losses. In turn, both sides become risk ac-ceptant with regard to the escalation or militarization of the conflict. Thus, thereis increased willingness and an increased probability of escalation to militarizedconflict.

Conclusion: An Agenda for Future Research

In sum, political geography has clearly made contributions to the study of inter-national conflict within the broader context of the study of international relations.Various geopolitical linkages have demonstrated the utility of thinking in terms ofspatiality, distance and proximity, and territory. Students of international relationshave drawn upon the work of geographers in this cumulative endeavor, but wehave not gone far enough. One important aspect of a future agenda for the study ofinternational politics and international conflict is to exploit more fully the waysgeography can help us think about international phenomena and design researchaccordingly.

There are a number of new (and exciting) developments in geography that caninform the study of international conflict. These developments stem from theoret-ical and methodological work by geographers that parallels debates and advances inthe study of international relations. Philosophically, geographers are paying greaterattention to what ‘‘spatial reality’’ is, what it means, and the multiple ways in whichscholars approach the geographic world. These are questions of geospatial ontologythat are similar to those scholars of international relations would recognize as beingraised by constructivist approaches to social scientific investigation, for example, the

HARVEY STARR 399

work of Barry Smith (2001; Smith and Barzi 2000). Galton (2001:173) gets to theheart of the matter quickly and simply:

Everything I see on a map can be described as geographical information. It isobvious that such information comes in many different forms. Representing atown by placing a small circle at a specific location on the map is quite differentfrom showing the extent of woodland by coloring areas of the map green. Howshould the different kinds of geographical information be classified? Can wedivide them into a small number of basic kinds, or are we faced with a plethora ofuniquely different sorts of information that resists any attempt at systematization?If we settle for a few basic kinds, how are these kinds related, and how should wedecide which kind to use in any particular case?

What this quote tells us is that we must revisit our traditional measures of prox-imity or ‘‘distance,’’ recognizing explicitly the psychological aspects of place, dis-tance, their meanings, and how those meanings can change. We need to extendpolitical psychology to concerns that we might have previously called political ge-ography, and vice versa. Thus, we must continue to stress that such meanings aredynamic and that geographic contexts are not immutable constant factors in ourresearch. Such additional approaches can have a significant impact on studies ofethnic conflict (for example, Lake and Rothchild 1998) and especially on the studyof protracted social conflict. In protracted social conflicts both sides usually haveinvested territory with great importance for the identity of the group, therebymaking compromise over territorial adjustments difficult. Thus, one challenge willbe not only to look at the presence or absence of contiguity (especially as a ‘‘control’’variable; see Ray 2003), but how territory is viewed by leaders, populations, andrelevant subsets of these populations. The spatial distribution and locations of suchsubsets of people should be examined as well.

We also need to pay explicit attention to geographically based or territoriallyoriented identities. Our research agenda must include ethnic-based (or national)identities that are commonly used (for example, as in Gurr’s 1993 research onendangered minorities). We presently have studies of territorial claims, territorialtransfers, and territorial conflict. Although any study of international conflict needsto be concerned with how territory might be involved in the stakes or causes ofconflict, territory also needs to be factored explicitly into studies that use ethnic orcommunal identity as variables.

At the same time, as briefly outlined in Starr (2001a, 2003), a number of meth-odological innovations are taking place in the study of the spatial context of politicalphenomena that have been developed by or borrowed from geographers. Suchdevelopments now allow students of international conflict to add spatially basedanalyses to our more standard time-based research designs. Indeed, one of themost important components of a future research agenda is the use of spatial sta-tistics.

Anselin (1999:93) has noted three major ways in which the methodologies ofspatial analysis contribute to the social scientist’s ‘‘toolbox.’’ The first deals with‘‘data integration,’’ which involves ‘‘the conversion of data collected at one spatialscale (and time dimension) to other scales and dimensions.’’ This need has gen-erated the increased use of GIS or Geographic Information Systems across thesocial sciences and has been driven by the increased availability of spatial dataproduced by such systems. Anselin (1995:93) argues that GISFwith its ability togenerate and analyze large datasets, along with its other organizational, modeling,and synthesizing attributesFhas ‘‘created a demand for new techniques for spatialdata analysis of both an exploratory and a confirmatory nature.’’ In part based onthe nature and power of GIS, a second important contribution involves ‘‘explor-atory spatial data analysis (ESDA) and visualization in an inductive approach todiscovering patterns, eliciting hypotheses, and suggesting associations’’ (Anselin


1999:68). A third important contribution focuses on the use of deductive ap-proaches to analyze the spatial context, specifically the use of the specialized meth-odology of spatial statistics and spatial econometrics. An overview is provided byMichael Goodchild et al. (2000:148):

Principles of statistics in a spatial context are contained in spatial statistics, geo-statistics, and spatial econometrics and based on the concept of a spatial randomfield. They incorporate models of spatial dependence and spatial heterogeneity,allowing for specification testing, estimation, and prediction of spatial phenomenaobserved as points, continuous surfaces, or lattices (regions). Spatial data analysistools therefore permit exploration of data from a spatial perspective, looking forspatial patterns, correlations, outliers, and residuals and submitting apparentpatterns to rigorous statistical tests. They also permit the confirmatory testing ofnonspatial hypotheses using spatial data.

Political geography, then, also provides the possibility for a variety of new datasets for the study of international phenomena. Looking at space as well as varioussubdivisions of that space is one strategy by which to increase the number of ob-servations in what have previously been small-N studies (for example, see Collier1993). GIS have also generated a large amount of new data about the world andhow it is organized. Part of our agenda is to identify such data and make use of it.Another part of the agenda is to design new GIS-based studies to develop data setsthat would have been unavailable without GIS technology.

One simple definition of what a geographic information system includes is pro-vided by geographer Jack Dangermond (1992:11–12): ‘‘a GIS is an organized col-lection of computer hardware, software, and geographic data designed to efficientlycapture, store, update, manipulate, analyze, and display all forms of geographicallyreferenced information.’’ It is important to understand that a GIS is a tool foundedon a variety of computer technologies that permits the handling of data concernedwith the locationFor, more broadly, the spatialityFof physical phenomena andhuman artifacts. A GIS permits the integration of data about the spatiality of phe-nomena along with data about other characteristics of those phenomena. Moretechnically (Dangermond 1992:12):

GISs can store geographically referenced (cartographic or spatial) data in a raster(grid or cellular-based) data structure or in an x,y coordinate reference-based(vector) data structure as points (nodes), lines (arcs), and polygons (bounded byarcs, inclosing an area). . . . GISs make use of a variety of coordinate referencingsystems to locate features on the earth relative to others; these coordinate sys-tems, in turn, make use of a variety of map projections to transform earth ref-erences onto a two-dimensional surface (the map).

It is important to see that GIS goes well beyond mere computer mapping. Ac-cording to David Cowen (1990:57), the heart of a GIS system is its ability to overlayvarious layers or coverages of data and in the process to create ‘‘new informationrather than just have retrieved previously encoded information.’’ To do so, a GISsystem must include the following four major components (Marble 1990:10):

(1) A data input subsystem that collects or processes spatial data derived fromexisting maps, remote sensors, and so on;

(2) A data storage and retrieval subsystem that organizes the spatial data in aform that permits it to be quickly retrieved by the user for subsequentanalysis;

(3) A data manipulation and analysis subsystem that performs a variety oftasks such as changing the form of the data through user-defined aggre-gation rules or producing estimates of parameters and constraints forvarious space-time organization or simulation models;

HARVEY STARR 401

(4) A data reporting subsystem that is capable of displaying all or part of theoriginal database as well as manipulated data and the output from spatialmodels in tabular or map form.

More international relations scholars are taking advantage of the data and meth-ods of geographic information systems. Such studies have been directed to bothgeography-territory as the source of conflict as well as geography-territory as fa-cilitating conditions. As an example of the former, Paivi Lujala and Halvard Buhaug(2003) use a GIS to look at the spatial distribution of the assets of a territory inorder to link them to the study of armed conflict. By using GIS, Lujala and Buhaugcan do analyses at substate levels, disaggregate country-level data and analyses, andincrease the number of units studied.

Currently, one of the more extensive uses of geographic information systems hasbeen by the author (for example, Starr 2001b, 2002b; Starr and Thomas 2002a,2005). GIS-generated data have permitted a reconceptualization of borders, alongwith a new approach to operationalizing opportunity and willingness (see Starr2001a, 2002a for descriptions of the data set). Looking at GIS-generated indices for‘‘ease of interaction’’ and ‘‘salience’’ that can apply to any single border or segmentof any border, this project has revisited geography as a facilitating factor (oppor-tunity as ease of interaction) and geography as a source of conflict (by looking atwillingness to engage in conflict or cooperation through the importance or ‘‘sa-lience’’ of any border area). The research findings of Starr and Thomas (2005) onthe curvilinear relationship between contiguity and conflict noted above could nothave been discovered (or studied!) using previously collected data on contiguity, inwhich contiguity was an ‘‘on-off ’’ variable simply noting whether or not two statesshared a land border. This research and its results were only possible with the typeof data provided by GIS regarding the nature of borders. Thus, a future researchagenda should include revisiting studies that have used the simple presence orabsence of contiguity with data that now can detail the ‘‘nature’’ of specific bordersor portions of those borders. For example, which types of borders are most or leastrelated to spatial diffusion: those with high ease of interaction, high salience, orboth (‘‘vital borders’’)? Is it simply ‘‘borderness’’ in the sense of contiguity, orcloseness in miles or kilometers; or are there more specific qualities about thenature of the borders that are involved? Indeed, is it the nature of the borderrather than the numbers of borders that affect conflict behavior? What sorts ofborders can be found between states in enduring rivalries? What is the nature of theterritory over which conflicts arise in terms of importance or salience and ease ofinteraction?

As Starr (2003:370) has observed: ‘‘Cumulation and ‘progress’ in the study ofglobal phenomena will depend on the quality and rigor of our theories and ourmethods. Synthesis will follow broad agent-structure approaches that cut acrossmore standard levels of analysis and disciplinary boundaries. The challenges facingresearchers arise from finding the appropriate methods by which to study theagent-structure problem.’’ The challenges listed in Starr (2002a) include the ques-tions of how to cut into the continuous feedback loops between agent and structureand between endogenous and exogenous factors, how to design studies of necessity,and how to craft the proper designs and methods to study substitutability. Anadditional challenge we face is to increase the relevance of the agent-structureproblem (Starr 2003). For research on international conflict to meet the stringentcriteria of integrative cumulation, we will also need to meet the challenge of how tostudy space and spatiality and the even more difficult challenge of combining spa-tial and temporal contexts, perspectives, and modes of analysis. This last challengeis both the motivation for the current essay and the main target of a future agendafor research.


References

ABLER, RONALD, JOHN S. ADAMS, AND PETER GOULD. (1971) Spatial Organization: The Geographer’s Viewof the World. Englewood Cliffs, NJ: Prentice-Hall.

AGNEW, JOHN A., AND STUART CORBRIDGE. (1995) Mastering Space. London: Routledge.AGNEW, JOHN. A., AND JAMES S. DUNCAN. (1989) Introduction. In The Power of Place, edited by John A.

Agnew and James S. Duncan. Boston: Unwin Hyman.ANSELIN, LUC. (1995) Local Indicators of Spatial AssociationFLISA. Geographical Analysis 27:

93–115.ANSELIN, LUC. (1999) The Future of Spatial Analysis in the Social Sciences. Geographic Information

Sciences 5:67–76.AXELROD, ROBERT, AND D. SCOTT BENNETT. (1993) A Landscape Theory of Aggregation. British Journal

of Political Science 23:211–233.AZAR, EDWARD E. (1984) The Theory of Protracted Social Conflict and the Challenge of Transforming

Conflict Situations. In Conflict Processes and the Breakdown of International Systems, edited by DinaZinnes. Denver: Monograph Series in World Affairs, University of Denver.

AZAR, EDWARD E. (1985) Protracted International Conflicts: Ten Propositions. International Interactions12(1):59–70.

BOULDING, KENNETH E. (1962) Conflict and Defense. New York: Harper and Row.BREMER, STUART A. (1992) Dangerous Dyads: Conditions Affecting the Likelihood of Interstate War,

1816–1965. Journal of Conflict Resolution 36:309–341.BUENO DE MESQUITA, BRUCE. (1981) The War Trap. New Haven: Yale University Press.BUENO DE MESQUITA, BRUCE. (1989) The Contribution of Expected-Utility Theory to the Study of

International Conflict. In Handbook of War Studies, edited by Manus I. Midlarsky. Boston: UnwinHyman.

CIOFFI-REVILLA, CLAUDIO, AND HARVEY STARR. (1995) Opportunity, Willingness, and Political Uncer-tainty: Theoretical Foundations of Politics. Journal of Theoretical Politics 7(4):447–476.

COLLIER, DAVID. (1993) The Comparative Method. In Political Science: The State of the Disci-pline II, edited by Ada W. Finifter. Washington, DC: The American Political Science As-sociation.

COWEN, DAVID J. (1990) GIS versus CAD versus DBMS: What are the Differences. In IntroductoryReadings in Geographic Information Systems, edited by Donna J. Peuquet and Duane F. Marble.London: Taylor & Francis.

DANGERMOND, JACK. (1992) What Is a Geographic Information System (GIS)? In GeographicInformation Systems (GIS) and MappingFPractices and Standards, edited by A. I. Johnson, C. B.Pettersson, and J. L. Fulton. Philadelphia: ASTM.

DIEHL, PAUL F. (1991) Geography and War: A Review and Assessment of the Empirical Literature.International Interactions 17:11–27.

DIEHL, PAUL F., ED. (1999) A Road Map to War. Nashville: Vanderbilt University Press.DODDS, KLAUS. (2001) Political Geography III: Critical Geopolitics after Ten Years. Progress in Human

Geography 25(3):469–484.ENTERLINE, ANDREW J. (1998) Regime Changes, Neighborhoods, and Interstate Conflict, 1816–1992.

Journal of Conflict Resolution 42(6):804–829.ERIKSSON, MIKAEL, AND PETER WALLENSTEEN. (2004) Armed Conflict, 1989–2003. Journal of Peace

Research 41(5):625–636.FRIEDMAN, GIL. (1999) Conceptualizing Protracted Conflict and Protracted Conflict Management.

In The Understanding and Management of Global Violence, edited by Harvey Starr. New York:St. Martin’s Press.

FRIEDMAN, GIL. (2002) Toward A Spatial Model of Protracted Conflict Management: The PalestinianCase. Unpublished Doctoral Dissertation, University of South Carolina.

GALTON, ANTONY. (2001) Space, Time, and the Representation of Geographical Reality. Topoi 20:173–187.

GIBLER, DOUGLAS M. (1996) Alliances That Never Balance: The Territorial Settlement Treaty. ConflictManagement and Peace Science 15(1):75–97.

GIBLER, DOUGLAS M., AND JOHN A. VASQUEZ. (1998) Uncovering the Dangerous Alliances, 1495–1980.International Studies Quarterly 42:785–807.

GLEDITSCH, KRISTIAN S. (2002) All International Politics is Local. Ann Arbor: University of MichiganPress.

GLEDITSCH, KRISTIAN S., AND MICHAEL D. WARD. (2001) Measuring Space: A Minimum DistanceDatabase. Journal of Peace Research 38(6):749–768.

HARVEY STARR 403

GOCHMAN, CHARLES S. (1992) Interstate Metrics: Conceptualizing, Operationalizing, and Measuringthe Geographic Proximity of States Since the Congress of Vienna. In The New Geopolitics, editedby Michael D. Ward. Philadelphia: Gordon and Breach.

GOERTZ, GARY. (1992) Contextual Theories and Indicators in World Politics. International Interactions17(4):285–303.

GOERTZ, GARY. (1994) Contexts of International Politics. Cambridge: Cambridge University Press.GOERTZ, GARY, AND PAUL F. DIEHL. (1992) Territorial Changes and International Conflict. London: Rout-

ledge.GOODCHILD, MICHAEL F., LUC ANSELIN, RICHARD P. APPELBAUM, AND BARBARA HERR HARTHORN.

(2000) Toward Spatially Integrated Social Science. International Regional Science Review 23:139–159.

GURR, TED ROBERT. (1993) Minorities at Risk. Washington, DC: United States Institute of Peace Press.HAMMARSTROM, MATS, AND BIRGER HELDT. (2002) The Diffusion of Military Intervention: Testing a

Network Position Approach. International Interactions 28(4):355–377.HASSNER, RON E., AND ANN HIRONAKA. (2002) Can Time Heal All Wounds? Findings From Territorial

Disputes. Paper presented at the annual meeting of the International Studies Association, NewOrleans, March.

HENDERSON, ERROL. (1997) Culture or Contiguity: Ethnic Conflict, the Similarity of States, and theOnset of War, 1820–1989. Journal of Conflict Resolution 41(5):649–668.

HENSEL, PAUL R. (1996) Charting a Course to Conflict: Territorial Issues and Militarized InterstateDisputes, 1816–1992. Conflict Management and Peace Science 15(1):43–73.

HENSEL, PAUL R. (2000) Territory: Theory and Evidence on Geography and Conflict. In What Do WeKnow About War?, edited by John A. Vasquez. Boulder: Rowman and Littlefield.

HENSEL, PAUL R. (2001) Contentious Issues and World Politics: The Management of Territorial Claimsin the Americas, 1816–1992. International Studies Quarterly 45:81–109.

HENRIKSON, ALAN K. (2002) Distance and Foreign Policy: A Political Geography Approach. Interna-tional Political Science Review 23(4):437–466.

HOLSTI, K. J. (1991) Peace and War: Armed Conflicts and International Order, 1648–1989. Cambridge:Cambridge University Press.

HUTH, PAUL K. (1996) Standing Your Ground: Territorial Disputes and International Conflict. Ann Arbor:University of Michigan Press.

HUTH, PAUL K., AND TODD L. ALLEE. (2002) The Democratic Peace and Territorial Conflict in the TwentiethCentury. Ann Arbor: University of Michigan Press.

KACOWICZ, ARIE M. (1994) The Problem of Peaceful Territorial Change. International Studies Quarterly38:219–254.

KIRBY, ANDREW, AND MICHAEL D. WARD. (1987) Space, Spatiality, Geography, Territoriality, Context,LocaleFand Conflict. Paper presented at the annual meeting of the American Political ScienceAssociation, Chicago, August.

KLIOT, NURIT, AND DAVID NEWMAN, EDS. (2000) Geopolitics at the End of the Twentieth Century. London:Frank Cass.

KOCS, STEPHEN. (1995) Territorial Disputes and Interstate War, 1945–1987. Journal of Politics 57:159–175.

LAKE, DAVID A., AND DONALD ROTHCHILD, EDS. (1998) The International Spread of Ethnic Conflict.Princeton: Princeton University Press.

LEEDS, BRETT ASHLEY, AND DAVID R. DAVIS. (1999) Beneath the Surface: Regime Type and Inter-national Interaction, 1953–1978. Journal of Peace Research 36:5–21.

LEMKE, DOUGLAS. (1995) The Tyranny of Distance: Redefining Relevant Dyads. International Inter-actions 21(1):23–38.

LEMKE, DOUGLAS, AND WILLIAM REED. (2001) The Relevance of Politically Relevant Dyads. Journal ofConflict Resolution 45(1):126–144.

LEVY, JACK S. (2000) Loss Aversion, Framing Effects, and International Conflict: Perspectives fromProspect Theory. In Handbook of War Studies II, edited by Manus I. Midlarsky. Ann Arbor:University of Michigan Press.

LUJALA, PAIVI, AND HALVARD BUHAUG. (2003) How to Deal with the Spatial Dimension: GIS Data onIntrastate Conflict and Natural Resources. Paper presented at the annual meeting of the In-ternational Studies Association, Portland, February.

MAOZ, ZEEV. (1996) Domestic Sources of Global Change. Ann Arbor: University of Michigan Press.MAOZ, ZEEV. (2001) Democratic Networks: Connecting National, Dyadic, and Systemic Levels of

Analysis in the Study of Democracy and War. In War in a Changing World, edited by Zeev Maozand Azar Gat. Ann Arbor: University of Michigan Press.


MAOZ, ZEEV, AND BRUCE RUSSETT. (1992) Alliance, Contiguity, Wealth, and Political Stability: Is theLack of Conflict Among Democracies a Statistical Artifact? International Interactions 17(3):245–267.

MARBLE, DUANE F. (1990) Geographic Information Systems: An Overview. In Introductory Readings inGeographic Information Systems, edited by Donna J. Peuquet and Duane F. Marble. London: Taylor& Francis.

MITCHELL, SARA MCLAUGHLIN, AND BRANDON C. PRINS. (1999) Beyond Territorial Contiguity: Issuesat Stake in Democratic Militarized Interstate Disputes. International Studies Quarterly 43:169–183.

MORTON, JEFFREY S., AND HARVEY STARR. (2001) Uncertainty, Change, and War: Power Fluctuationsand War in the Modern Elite Power System. Journal of Peace Research 38:49–66.

MOST, BENJAMIN A., AND HARVEY STARR. (1980) Diffusion, Reinforcement, Geopolitics, and the Spreadof War. American Political Science Review 74:932–946.

MOST, BENJAMIN A., AND HARVEY STARR. (1989) Inquiry, Logic, and International Politics. Columbia:University of South Carolina Press.

MOST, BENJAMIN, HARVEY STARR, AND RANDOLPH SIVERSON. (1989) The Logic and Study of theDiffusion of International Conflict. In Handbook of War Studies, edited by Manus Midlarsky.Boston: Unwin Hyman.

MURDOCH, JAMES C., AND TODD SANDLER. (2002) Economic Growth, Civil Wars, and Spatial Spillovers.Journal of Conflict Resolution 46:91–110.

NEWMAN, DAVID. (1999) Real Spaces, Symbolic Spaces: Interrelated Notions of Territory in the Arab-Israeli Conflict. In A Road Map to War, edited by Paul Diehl. Nashville: Vanderbilt UniversityPress.

O’LOUGHLIN, JOHN, AND LUC ANSELIN. (1991) Bringing Geography Back to the Study of InternationalRelations: Spatial Dependence and Regional Context in Africa, 1966–1978. International Inter-actions 17(1):29–61.

O’LOUGHLIN, JOHN, AND LUC ANSELIN. (1992) Geography of International Conflict and Cooperation:Theory and Methods. In The New Geopolitics, edited by Michael D. Ward. Philadelphia: Gordonand Breach.

PRUITT, DEAN G., AND SUNG HEE KIM. (2004) Social Conflict: Escalation, Stalemate, and Settlement. Boston:McGraw-Hill.

RASLER, KAREN, AND WILLIAM R. THOMPSON. (2000) Explaining Rivalry Escalation to War: Space,Position, and Contiguity in the Major Power System. International Studies Quarterly 44:503–530.

RAY, JAMES L. (2003) Explaining Interstate Conflict and War: What Should Be Controlled For? ConflictManagement and Peace Science 20:1–31.

RUSSETT, BRUCE M. (1967) International Regions and the International System. Chicago: Rand McNally.RUSSETT, BRUCE M., AND HARVEY STARR. (2000) From Democratic Peace to Kantian Peace: Democracy

and Conflict in the International System. In Handbook of War Studies II, edited by Manus I.Midlarsky. Ann Arbor: University of Michigan Press.

SAMPLE, SUSAN G. (2002) The Outcomes of Military Buildups: Minor States vs. Major Powers. Journalof Peace Research 39(6):669–691.

SIMOWITZ, ROSLYN. (1998) Evaluating Conflict Research on the Diffusion of War. Journal of PeaceResearch 35:211–230.

SIVERSON, RANDOLPH M., AND HARVEY STARR. (1991) The Diffusion of War. Ann Arbor: University ofMichigan Press.

SMITH, BARRY. (2001) Fiat Objects. Topoi 20:131–148.SMITH, BARRY, AND A. C. VARZI. (2000) Fiat and Bona Fide Boundaries. Philosophy and Phenomenological

Research 60:401–420.SPROUT, HAROLD, AND MARGARET SPROUT. (1969) Environmental Factors in the Study of International

Politics. In International Politics and Foreign Policy, edited by James N. Rosenau. New York: FreePress.

STARR, HARVEY. (1978) ‘‘Opportunity’’ and ‘‘Willingness’’ as Ordering Concepts in the Study of War.International Interactions 4:363–387.

STARR, HARVEY. (2001a) The Power of Place and the Future of Spatial Analysis in the Study of Conflict.Presidential Address presented at the 35th North American Meeting of the Peace Science Society,Atlanta, October.

STARR, HARVEY. (2001b) Reconceptualizing International Borders Through the Application of GIS(Geographic Information Systems): A New Dataset on the ‘‘Nature’’ of Borders. InternationalStudies Perspectives 2(4):441–443.

HARVEY STARR 405

STARR, HARVEY. (2002a) Cumulation, Synthesis, and Research Design for the ‘‘Post-Fourth Wave’’. InMillennium Reflections on International Studies, edited by Michael Brecher and Frank Harvey. AnnArbor: University of Michigan Press.

STARR, HARVEY. (2002b) Opportunity, Willingness, and Geographic Information Systems: Reconcep-tualizing Borders in International Relations. Political Geography 21:243–261.

STARR, HARVEY. (2002c) Visions of Global Politics as an Intellectual Enterprise: Three QuestionsWithout Answers. In Visions of International Relations: Assessing an Academic Field, edited by DonaldPuchala. Columbia: University of South Carolina Press.

STARR, HARVEY. (2003) The Power of Place and the Future of Spatial Analysis in the Study of Conflict.Conflict Management and Peace Science 20(1):1–20.

STARR, HARVEY, AND BENJAMIN A. MOST. (1976) The Substance and Study of Borders in InternationalRelations Research. International Studies Quarterly 20:581–620.

STARR, HARVEY, AND RANDOLPH M. SIVERSON. (1990) Alliances and Geopolitics. Political GeographyQuarterly 9:232–248.

STARR, HARVEY, AND RANDOLPH M. SIVERSON. (1998) Cumulation, Evaluation, and the ResearchProcess: Investigating the Diffusion of Conflict. Journal of Peace Research 35:231–237.

STARR, HARVEY, AND G. DALE THOMAS. (2002a) The ‘‘Nature’’ of Contiguous Borders: Ease ofInteraction, Salience, and the Analysis of Crisis. International Interactions 28(3):213–235.

STARR, HARVEY, AND G. DALE THOMAS. (2002b) ‘‘Distance,’’ Interaction, and ‘‘Events’’: The Nature ofBorders and the Diffusion of Democracy. Paper delivered at 36th North American Meeting ofthe Peace Science Society (International), Tucson, October.

STARR, HARVEY, AND G. DALE THOMAS. (2005) The Nature of Borders and International Conflict:Revisiting Hypotheses on Territory. International Studies Quarterly 49:123–139.

SYMPOSIUM. (1985) Methodological Foundations of the Study of International Conflict. InternationalStudies Quarterly 29(2).

THIES, CAMERON G. (2001) Territorial Nationalism in Spatial Rivalries. International Interactions27(4):399–441.

VANZO, JOHN P. (1999) Border Configuration and Conflict: Geographical Compactness as a TerritorialAmbition of States. In A Road Map to War, edited by Paul Diehl. Nashville: Vanderbilt UniversityPress.

VASQUEZ, JOHN A. (1993) The War Puzzle. Cambridge: Cambridge University Press.VASQUEZ, JOHN A. (1996) Distinguishing Rivals That Go to War from Those That Do Not: A Quan-

titative Comparative Case Study of the Two Paths to War. International Studies Quarterly 40:531–558.

VASQUEZ, JOHN A., AND PAUL D. SENESE. (2003) A Unified Explanation of Territorial Conflict: Testingthe Impact of Sampling Bias, 1919–1992. International Studies Quarterly 47:275–298.

WARD, MICHAEL DON, ED. (1992) The New Geopolitics. Philadelphia: Gordon and Breach.WRIGHT, QUINCY. (1942) A Study of War. Chicago: University of Chicago Press.ZINNES, DINA A. (1976) The Problem of Cumulation. In In Search of Global Patterns, edited by James N.

Rosenau. New York: Free Press.ZIPF, G. K. (1949) Human Behavior and the Principle of Least Effort. Cambridge, MA: Addison-Wesley.


territory, proximity, and spatiality: the geography of international

Documents